 |
Highly available and scalable platforms are important for users of all contact center technology platforms, and are an absolute requirement for service providers offering contact center technology as a hosted service to their customers. CosmoCall Universe uses a combination of n+1 redundancy and clustering technology to achieve high scalability and fault tolerance to accommodate the Service Provider’s requirements.
The CosmoCall Universe Call Management Server (CMS) is one of the Core ACD Servers, and is made fault tolerant by implementing clustering technology so that there is always a backup if something should fail. Even in the event that one of the CMS machines should fail, the rest of the system will continue to handle the workload and process the calls as normal. No calls are lost, existing calls continue to be handled, and arriving calls continue to be accepted into the system. The CMS is a signaling point in the system and does not handle call media, so it requires minimal CPU cycles for each call it handles and thus is highly scalable.
The remaining server components are made scalable and fault tolerant using the n+1 redundancy model so that there are always enough servers to handle the workload if a server should fail. Standard load-balancing tools are used to evenly spread traffic across these servers and ensure that arriving calls continue to be handled gracefully even in the event of a failure. In an n+1 configuration, a single server failure will reduce the capacity to what can be handled by n servers. In a properly designed system, n servers will be capable of handling 100% of rated capacity. Therefore a single failure will not affect the rated system capacity.
The use of n+1 for high availability explains not only the high availability strategy, but also the scalability strategy. All of the n+1 components operate independently, and there is no theoretical limit on how many of each can be deployed in one system. The single system image is maintained by the Core ACD Servers. The current system supports 8,000 or more simultaneous active agents, 50,000 simultaneous calls and messages in the system, and over 500,000 busy hours, with development underway to take the system to significantly higher capacities.
|
|
